We noted in Chapter 17 that most stars are not single but instead are members of binary systems. 
(Sec. 17.9) Although many pulsars are known to be isolated (that is, not part of any binary), there is strong evidence that at least some do have binary companions, and the same is true of neutron stars in general (that is, even the ones not seen as pulsars).
The late 1970s saw several important discoveries about neutron stars in binary-star systems. Numerous X-ray sources were discovered near the central regions of our galaxy and also near the centers of a few rich star clusters. Some of these sources, known as X-ray bursters, emit much of their energy in violent eruptions, each thousands of times more luminous than our Sun, but lasting only a few seconds. A typical burst is shown in Figure 22.5.
Figure 22.5 An X-ray burster produces a sudden, intense flash of X-rays, followed by a period of relative inactivity lasting as long as several hours. Then another burst occurs. The bursts are thought to be caused by explosive nuclear burning on the surface of an accreting neutron star, similar to the explosions on a white dwarf that give rise to novae. (a) An optical photograph of the star cluster Terzan 2, showing a 2" dot at the center where the X-ray bursts originate. (b) X-ray images taken before and during the outburst. The most intense X-rays correspond to the position of the dot shown in frame (a).
This X-ray emission is thought to arise on or near neutron stars that are members of binary systems. Matter torn from the surface of the (main-sequence or giant) companion by the neutron star's strong gravitational pull accumulates on the neutron star's surface. As in the case of white-dwarf accretion (see Chapter 21), the material does not fall directly onto the surface. Instead, as illustrated in Figure 22.6(a), it forms an accretion disk. (Compare with Figure 21.2, which depicts the white-dwarf equivalent.) The gas goes into a tight orbit around the neutron star, then slowly spirals inward. The inner portions of the accretion disk become extremely hot, releasing a steady stream of X-rays.
Figure 22.6 (a) Matter flows from a normal star toward a compact neutron-star companion and falls toward the surface in an accretion disk. As the gas spirals inward under the neutron star's intense gravity it heats up, becoming so hot that it emits X-rays. In at least one instance—the peculiar object SS433—some material may be ejected in the form of two high-speed jets of gas. (b) False-color radiographs of SS433, made at monthly intervals (left to right), show the jets moving outward and the central source rotating under the gravitational influence of the companion star.
As gas builds up on the neutron star's surface its temperature rises due to the pressure of overlying material. Eventually, it becomes hot enough to fuse hydrogen. The result is a sudden period of rapid nuclear burning that releases a huge amount of energy in a brief but intense flash of X-rays—an X-ray burst. After several hours of renewed accumulation, a fresh layer of matter produces the next burst. Thus, an X-ray burst is much like a nova explosion on a white dwarf, but occurring on a far more violent scale because of the neutron star's much stronger gravity. (Sec. 21.1)
Not all the infalling gas makes it onto the neutron star surface, however. In at least one case—an object known as SS433*—we have direct observational evidence that some material is instead shot completely out of the system at enormously high speeds. SS433 expels more than one Earth mass of material every year in the form of two oppositely directed narrow jets moving roughly perpendicular to the disk. Observations of the Doppler shifts of optical emission lines produced within the jets themselves imply speeds of almost 80,000 km/s—over 25 percent of the speed of light! As the jets interact with the interstellar medium they emit radio radiation, as shown in Figure 22.6(b).
*(The name simply identifies it as the 433rd entry in a particular catalog of stars with strong optical emission lines.)
Jets of this sort are apparently quite common in astronomical systems in which an accretion disk surrounds a compact object (such as a neutron star or a black hole). They are believed to be produced by the intense radiation and magnetic fields near the inner edge of the disk, although the details of their formation are still uncertain. Incidentally, they are not the "lighthouse" beams of radiation from the neutron star itself (Figure 22.4) that can result in a pulsar. Although SS433 is the only stellar object currently known to produce jets, we will see examples of similar phenomena on much larger scales in later chapters. One of the most important aspects of SS433 is that we can actually study both the disk and the jets instead of simply having to assume their existence, as in more distant cosmic objects.
Discovered serendipitously in the late 1960s by military satellites looking for violators of the Nuclear Test Ban Treaty, and first made public in the 1970s, gamma-ray bursts have developed into one of the deepest mysteries in astronomy today. Observationally, the bursts consist of bright, irregular flashes of gamma rays typically lasting only a few seconds, with substantial variations from one observed burst to another (Figure 22.7a). Until recently, it was thought that gamma-ray bursts were basically "scaled-up" versions of X-ray bursters, in which matter accreted from the binary companion experienced even more violent nuclear burning, accompanied by the release of the more energetic gamma rays. However, it now appears that this is not the case.
Figure 22.7 (a) Plots of intensity versus time (in seconds) for some gamma-ray bursts. Note the substantial differences between them. Some bursts are irregular and spiky, whereas others are much more smoothly varying. Whether this wide variation in burst appearance means that more than one physical process is at work is presently unknown. (b) Positions on the sky of the first 1000 gamma-ray bursts detected by GRO; BATSE stands for Burst and Transient Source Experiment. The bursts appear to be distributed isotropically (uniformly) across the entire sky.
Civilian satellites such as the Compton Gamma-Ray Observatory (GRO) detect gamma-ray bursts at the rate of about one a day. (Sec. 5.6) Figure 22.7(b) shows an all-sky plot of the positions of about a thousand bursts detected during the first three years of Compton's operation. Unlike similar maps at all wavelengths (such as Figure 5.34e), which generally show a distinct correlation with the plane of the Milky Way, the gamma-ray bursts plotted here are distributed randomly across the sky (their distribution is said to be "isotropic"). The bursts seemingly never repeat at the same location, show no obvious clustering, and appear unaligned with any known large-scale structure, near or far.
The isotropy of the GRO data has convinced most astronomers that the bursts must originate at so-called cosmological distances—that is, far beyond our Milky Way galaxy. In addition, Compton finds faint bursts to be less frequent than would be expected from an infinite, unbounded population of such sources. Apparently there is an outer limit of some sort in their distribution in space. However, it is unclear whether the bursts are associated with the isotropic spread of distant galaxies—or whether they are something else yet again, outside of galaxies and somewhere in intergalactic space.
If the bursts really do originate very far away in space, that must mean that their intrinsic power is enormous; otherwise they wouldn't be detectable by our equipment, even on-station above Earth's atmosphere. At cosmological distances, each burst must generate more energy than a typical supernova explosion—in only a few seconds!
Until 1997 there was one major problem with this line of reasoning: the actual distances to the gamma-ray bursts were completely unknown. As a result, the inference that the bursts were very distant, and hence very energetic, was indirect, based mainly on their isotropic distribution. The distances to the bursts were difficult to determine because astronomers were unable to find any optical counterparts or any kind of association with known cosmic objects. Part of the problem is that gamma rays are far too penetrating to be focused by conventional optics; consequently, GRO's burst positions are uncertain by several arc degrees.
All this changed on May 8, 1997, thanks to a remarkable combination of gamma-ray, X-ray, and optical observations of the gamma-ray burst GRB970508. The Italian—Dutch BeppoSAX satellite, launched in 1996, recorded the burst in both the gamma- and X-ray regions of the spectrum. The importance of the X-ray observations is that they allowed a much more accurate determination of the burst's location on the sky—in fact, to within a few arc minutes. That was enough for ground-based optical astronomers to look for and find GRB970508's optical afterglow. For the first time, astronomers had observed an optical counterpart to a gamma-ray burst.
The burst's optical spectrum, obtained using the Keck telescope, revealed a very important piece of information. Several absorption lines of iron and magnesium were identified, but they were redshifted by almost a factor of 2 in wavelength. Such redshifts, as we will see in Chapter 24, are the result of the expansion of the universe, and they are clear proof—a "smoking gun," if you will—that at least this gamma-ray burst, and presumably all others, really did occur at cosmological distances. The events responsible for GRB970508 apparently occurred more than 2000 megaparsecs from Earth. Figure 22.8 is an optical photograph of another gamma-ray burst, called GRB970228—one of the first such images that associates a gamma-ray burst with a potential host galaxy. In May 1998, astronomers announced a distance measurement for a second gamma-ray burst (GRB971214, which occurred on December 14, 1997), and again identified a host galaxy. According to the redshift of lines observed in that galaxy's spectrum, this burst was almost twice as far away as GRB970508, implying an enormous luminosity—perhaps as much as 100 times greater than that of a typical supernova. These observations lend strong support to the idea that gamma-ray bursts originate in faraway galaxies throughout the universe.
Figure 22.8 Optical photograph, in false color, of gamma-ray burst GRB970228. The optical counterpart of the gamma-ray burst (marked by the arrow) is a white blob to the upper left of image center; to the lower right is an extended object thought to be the burst's host galaxy.
Prior to the measurement of GRB970508's distance, some theorists had attempted to explain gamma-ray bursts in terms of nearby—and hence much less energetic—events, perhaps in the halo of our own galaxy (see Chapter 23) or even in the Oort cloud surrounding the Sun. (Sec. 14.2) However, it now seems clear that gamma-ray bursts are both very distant and extremely energetic. What's more, the millisecond flickering in the bursting gamma rays detected by Compton imply that, whatever their origin, all their energy must come from an extremely small volume—in fact, no larger than a few hundred kilometers across.
The reasoning is as follows. If the emitting region were, say, 300,000 km—1 light second—across, even an instantaneous change in intensity at the source would be smeared out over a time interval of 1 s as seen from Earth, because light from the far side of the object would take 1 s longer to reach us than light from the near side. For the gamma-ray variation not to be blurred by the light-travel time, the source cannot be more than 1 light millisecond, or 300 km, in diameter.
One leading mechanism for explaining gamma-ray bursts is the true end point of a binary-star system. Suppose that both members of the binary evolve to become neutron stars. As the system continues to evolve, gravitational radiation (see Interlude 22-1) is released and the two ultradense stars will spiral in toward each other. Once they are within a few kilometers of each another, coalescence is inevitable. Such a merger will likely produce a violent explosion, comparable in energy to a supernova, and this could conceivably explain the flashes of gamma rays we observe. However, the huge luminosity of the December 1997 burst may be too much even for this model to explain. Evidently, the final word on these enigmatic objects has yet to be written.
In the mid-1980s an important new category of pulsars was found—a class of very rapidly rotating objects called millisecond pulsars. Several dozen millisecond pulsars are currently known in the Milky Way galaxy. These objects spin hundreds of times per second (that is, their pulse period is a few milliseconds, 0.001 s). This speed is about as fast as a typical neutron star can spin without flying apart. In some cases, the star's equator is moving at more than 20 percent of the speed of light. This speed suggests a phenomenon bordering on the incredible—a cosmic object of kilometer dimensions, more massive than our Sun, spinning almost at breakup speed, making nearly 1000 complete revolutions every second. Yet the observations and their interpretation leave little room for doubt.
The story of these remarkable objects is further complicated because many of them are found in globular clusters. This is odd because globular clusters are known to be very old—10 billion years, at least. Yet Type II supernovae (the kind that create neutron stars) are associated with massive stars that explode within a few tens of millions of years after their formation, and no stars have formed in any globular cluster since the cluster itself came into being. Thus, no new neutron star has been produced in a globular cluster in a very long time. But, as earlier mentioned, the pulsar produced in a supernova explosion is expected to slow down in only a few million years. After 10 billion years its rotation should have all but ceased. The rapid rotation of the pulsars found in globular clusters cannot be a relic of their birth. These objects must have been spun up—that is, had their rotation rates increased—by some other, much more recent, mechanism.
The most likely explanation for the high rotation rate of these objects is that the neutron star has been spun up by drawing in matter from a companion star. As matter spirals down onto the star's surface in an accretion disk it provides the "push" needed to make the neutron star spin faster (see Figure 22.9). Theoretical calculations indicate that this process can spin the star up to breakup speed in about a hundred million years. Subsequently, an encounter with another star may eject the neutron star from the binary, or the pulsar's radiation may destroy its companion, so an isolated millisecond pulsar results. This general picture is supported by the finding that of the 40 or so millisecond pulsars seen in globular clusters, 10 are known currently to be members of binary systems. These numbers are quite consistent with the rate at which binaries can be broken up by encounters with other cluster members.
Figure 22.9 Gas from a companion star spirals down onto the surface of a neutron star. As the infalling matter strikes the star it moves almost parallel to the surface, so it tends to make the star spin faster. Eventually, this process can result in a millisecond pulsar—a neutron star spinning at the incredible rate of hundreds of revolutions per second.
Thus, although a pulsar like the Crab is the direct result of a supernova explosion, millisecond pulsars are the product of a two-stage process. The neutron star was formed in an ancient supernova, billions of years ago. Only relatively recently, through interaction with a binary companion, has it achieved the rapid spin that we observe today. Once again, we see how members of a binary system can evolve in ways quite different from single stars. Notice that the scenario of accretion onto a neutron star from a binary companion is the same scenario that we just used to explain the existence of X-ray bursters. In fact, the two phenomena are very closely linked. Many X-ray bursters may be on their way to becoming millisecond pulsars.
The way in which a neutron star can become a member of a binary system is the subject of active research, because the violence of a supernova explosion would be expected to blow the binary apart in many cases. Only if the supernova progenitor lost a lot of mass before the explosion would the binary be likely to survive. Alternatively, by interacting with an existing binary and displacing one of its components, a neutron star may become part of a binary system after it is formed, as depicted in Figure 22.10. Astronomers are eagerly searching the skies for more millisecond pulsars to test their ideas.
Figure 22.10 A neutron star can encounter a binary made up of two low-mass stars, ejecting one of them and taking its place. This mechanism provides a means of forming a binary system with a neutron-star component (which may later evolve into a millisecond pulsar) without having to explain how the binary survived the supernova explosion that formed the neutron star.
Radio astronomers can capitalize on the precision with which pulsar signals repeat themselves to make extremely accurate measurements of pulsar motion. In January 1992, radio astronomers at the Arecibo Observatory found that the pulse period of a recently discovered millisecond pulsar lying some 500 pc from Earth varies in an unexpected but quite regular way. Careful analysis of the data has revealed that the period fluctuates on two distinct time scales—one of 67 days, the other of 98 days. The changes in the pulse period are small—less than one part in 107—but repeated observations have confirmed their reality.
The leading explanation for these fluctuations holds that they are caused by the Doppler effect as the pulsar wobbles back and forth in space. But what causes the wobble? The Arecibo group believes it is the result of the combined gravitational pulls of not one but two planets, each about three times the mass of Earth! One orbits the pulsar at a distance of 0.4 A.U. and the other at a distance of 0.5 A.U. Their orbital periods are 67 and 98 days, respectively, matching the timing of the fluctuations. In April 1994 the group announced further observations that not only confirmed their earlier findings but also revealed the presence of a third body, with mass comparable to Earth's Moon, orbiting only 0.2 A.U. from the pulsar.
These remarkable results constitute the first definite evidence of planet-sized bodies outside our solar system. A few other millisecond pulsars have since been found with similar behavior. However, it is unlikely that any of these planets formed in the same way as our own. Any planetary system orbiting the pulsar's progenitor star was almost certainly destroyed in the supernova explosion that created the pulsar. As a result, scientists are still uncertain about how these planets came into being. One possibility involves the binary companion that provided the matter necessary to spin the pulsar up to millisecond speeds. Possibly the pulsar's intense radiation and strong gravity destroyed the companion, then spread its matter out into a disk (a little like the solar nebula) in whose cool outer regions the planets might have condensed.
Astronomers have been searching for decades for planets orbiting main-sequence stars like our Sun, on the assumption that planets are a natural by-product of star formation. (Sec. 15.2) Only very recently have these searches begun to yield positive results, although some of these findings remain controversial. (Interlude 15-1) It is ironic that the first (and only) Earth-sized planets to be found outside the solar system orbit a dead star and have little or nothing in common with our own world.
